Although numerous genetic alterations have been causally linked to cancer development, many of the affected genes appear to fall into common cellular pathways. With the advent of technologies capable of assessing the expression of thousands of genes simultaneously, the potential now exists to "reverse- engineer" cancer-related signaling pathways by computer modeling. Towards this end, the objective of this proposal is to develop experimental methods and analysis tools to elucidate the components of signaling networks functionally altered during tumorigenesis. To achieve this goal, prototype pathway modeling algorithms will be tested in Phase I for their ability to detect known relationships among genes manipulated in experimental systems. Specifically, recombinant wildtype versions of genes frequently mutated in cancer development will be inducibly expressed in cell lines, and the resulting expression changes in endogenous genes will be monitored by hybridization to oligonucleotide arrays. Analogous studies will also be performed using cells in which individual genes have been deleted. The protein products of genes that demonstrate altered expression in such studies serve as candidate functional mediators of the ectopically expressed genes. The commercial goal of these studies is to develop broadly applicable tools to facilitate the rational identification of novel diagnostic markers, prognostic markers, and therapeutic targets. PROPOSED COMMERCIAL APPLICATION: We will develop experimental and analytical tools to reverse-engineer signal transduction networks from gene expression profiles. These tools will facilitate the rational identification of novel diagnostic markers and therapeutic targets. As such, they will enhance the sale of high-throughput expression analysis systems and pathway analysis software.